PCA-Based Motion Energy Analysis

This capsule is a part of Behavior Video QC to apply Principal Component Analysis (PCA) to motion energy data extracted from Zarr files. The script loads motion energy data, applies PCA, generates spatial masks, and visualizes the results. Computing time depends on number and size of frames. An hour long video with ~600x400 frames at 60 Hz takes about 30 min.

Features

• Loads motion energy data from Zarr and NPZ files.
• Applies PCA to extract meaningful components from motion energy.
• Supports optional cropping of motion energy frames before PCA.
• To do: If cropping was applied during Motion Energy computation, loads 'cropepd_frames' indead of 'frames and skips the cropping stage. If crop is True, crops frames to using specified cropped_region.
• Generates and saves spatial masks for PCA components plots.
• Saves results, including PCA components, explained variance, and motion energy traces.
• Tracks processing time.

Prerequisites

Ensure you have the following dependencies installed:

pip install tqdm numpy zarr matplotlib

Usage

Running the script

Execute the script using:

python run_capsule.py

Parameters

Parameter	Description	Default Value
zarr_paths	List of paths to Zarr directories containing motion energy data.	utils.find_files()
npz_paths	List of paths to NPZ files storing motion energy traces.	utils.find_files()
crop	Boolean flag to enable cropping before PCA. Checks ME metadata.	None
crop_region	Tuple defining the crop region (y_start, x_start, y_end, x_end).	checks metadata
standardize4PCA	Boolean flag for standardizing data before PCA.	False
standardizeMasks	Boolean flag for standardizing spatial masks plots.	True

Example Output

Processing /root/capsule/data/video1.zarr
Applying PCA to motion energy data...
Saving PCA results...
Generating and saving spatial masks...
Plotting and saving PCA explained variance...
Plotting and saving motion energy trace...
Plotting and saving PCA component traces...
Total time taken: 360.45 seconds

Modifying Parameters

To process different Zarr and NPZ files, modify the root_dir parameter in script.py:

zarr_paths = utils.find_files(root_dir='/your/custom/path', endswith='zarr')
npz_paths = utils.find_files(root_dir='/your/custom/path', endswith='.npz', return_dir=False)

To change the cropping region, update the crop_region variable:

crop_region = (y_start, x_start, y_end, x_end)  # Adjust cropping dimensions in pixel values

To enable or disable standardization before PCA, update these parameters:

me_pca = PCAgenerator(zarr_path, npz_path, crop=True, crop_region=crop_region, 
                      standardize4PCA=True, standardizeMasks=False)